SavvySearch

» see savvysearch.com as it lived, on the wayback machine

~ the eulogy ~

SavvySearch went up in March 1995, written by Daniel Dreilinger while he was a graduate student in computer science at Colorado State University. The premise fit in a sentence. There were already dozens of search engines, no two of them covered the same part of the web, and querying all of them by hand was the actual labor of looking something up. The software did the querying and merged what came back.

The scoring is what made it a research project. SavvySearch kept a record of which engines had returned useful results for which kinds of query, and it watched how fast each one was answering at that moment, then used both to decide where to send the next question. Dreilinger and Adele Howe published the method in 1997, and that paper is why SavvySearch still turns up in metasearch literature long after every commercial rival of its generation dropped out of the citations.

By 1999 the front page had become a directory in its own right. You could metasearch the big general engines, or pick a specialty and metasearch only the sources that mattered there, so a tab for MP3s sat next to a tab for Usenet. CNET announced on 14 October 1999 that it was buying the company for 22 million dollars in cash and stock. By the following summer the brand was gone.

~ the dossier ~

~ snoop ~

Snoop was a button on the homepage that showed you a live feed of other people's queries. You clicked it and read what strangers had typed into the box a few seconds earlier, unedited. Several sites of the period ran a version of this and every one of them eventually switched it off, because the raw stream of what people ask a search engine when they believe nobody is watching is not something you want sitting beside a banner ad. It is the oldest ancestor of the query logs every search company now guards absolutely.

~ what i was using instead ~

I got online in 1998 on a 33.6k modem, and searching from Turkey in those years meant AltaVista, plus the Turkish directories for anything local. Metasearch was a category whose argument I understood and never adopted, because on a dial-up line the thing you feel is latency, and a tool that waits on six slow engines before it shows you anything feels worse than one engine that is merely mediocre. Google opened to the public in 1998 and by 2000 it was the default almost everywhere I looked. The problem metasearch existed to solve had shrunk to very little by the time CNET switched the domain over.

~ what CNET actually bought ~

CNET paid its 22 million for a company running SavvySearch.com plus two narrower metasearch sites, MP3meta.com and FinanceMeta.com. What it wanted was the technology for its own properties, and it already had a search brand in Search.com, which it had been operating since 1996. That accounts for the whole shutdown. The buyer owned a competing product with a better domain name and no reason to fund two front doors onto the same index.

The archive holds the changeover almost to the day. On 20 June 2000 www.savvysearch.com was still serving the yellow SavvySearch wordmark and the tagline about the search of all searches, with a small link reading "try Search.com!" already tucked above the query box. On 11 July 2000 the same address returned CNET Search.com. No capture from that summer contains a closure notice.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »